Böcke man, Pe i; Ko elainen, Mika; Salokangas, Hen i; Vaala uo, Ma ia
A icle — Published Ve sion
A amily a ai ? Long- e m economic and men al heal h
e ec s o spousal cance
Jou nal o Popula ion Economics
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: Böcke man, Pe i; Ko elainen, Mika; Salokangas, Hen i; Vaala uo, Ma ia
(2025) : A amily a ai ? Long- e m economic and men al heal h e ec s o spousal cance , Jou nal o
Popula ion Economics, ISSN 1432-1475, Sp inge , Be lin, Heidelbe g, Vol. 38, Iss. 1,
h ps://doi.o g/10.1007/s00148-025-01070-x
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/318427
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h p://c ea i ecommons.o g/licenses/by/4.0/
Jou nal o Popula ion Economics (2025) 38:19
h ps://doi.o g/10.1007/s00148-025-01070-x
ORIGINAL PAPER
A amily a ai ? Long- e m economic and men al heal h
e ec s o spousal cance
Pe i Böcke man1,2,3 ·Mika Ko elainen4,5 ·Hen i Salokangas5·
Ma ia Vaala uo5
Recei ed: 19 Ma ch 2023 / Accep ed: 2 Janua y 2025
© The Au ho (s) 2025
Abs ac
Resea ch on he amily spillo e e ec s o heal h shocks, which has ocused mainly
on labo ma ke ou comes, has yielded inconclusi e esul s, wi h limi ed insigh in o
long- e m consequences o unde lying mechanisms. We analyze he sho - and long-
e m impac s o cance on he una ec ed spouse’s labo supply and men al heal h
as well as ma i al s abili y, conside ing gende and ela i e income s a us wi hin he
households. Using popula ion-based egis e da a om Finland (1995–2019) and a
dynamic di e ence-in-di e ences design, we obse e wo key indings. Fi s , a can-
ce diagnosis leads o e y modes changes in a spouse’s labo supply bu signi ican
inc eases in he likelihood o psycho opic d ug use and psychia ic ou pa ien isi s.
Second, he main esul s mask conside able he e ogenei y ega ding ela i e income
wi hin he household. Seconda y ea ne s inc ease hei labo supply in esponse o
a al cance s bu dec ease i in non- a al cases, while b eadwinne s show small neg-
a i e esponses in bo h. Be ea ed women wi h lowe income sha e expe ience mo e
psychia ic symp oms, a end no obse ed in men. Ou indings e eal he impo ance
o p e-shock b eadwinne s a us in amily esponses o heal h shocks, sugges ing he
need o a ge ed suppo o ca egi ing and be ea ed spouses.
Keywo ds Heal h shock ·Cance ·Family spillo e e ec s ·Household di ision
o labo ·E en s udy ·Di e ence-in-di e ences ·Men al heal h ·Ma i al s abili y
JEL Classi ica ion I10 ·J12 ·J17 ·J22
1 In oduc ion
The e is a g owing body o esea ch on he indi ec e ec s o se e e heal h shocks
wi hin he amily (e.g., Ga cía-Gómez e al. 2013; Jeon and Pohl 2017; Fadlon and
Nielsen 2021; Vaala uo e al. 2023). Indi ec e ec s pe ain o amily membe s o he
Responsible edi o :XiChen
Ex ended au ho in o ma ion a ailable on he las page o he a icle
0123456789().: V,- ol 123
19 Page 2 o 30 P. Böcke man e al.
han he pe son who alls ill. Accu a ely quan i ying such e ec s, which ha e been
la gely neglec ed in he e alua ion o he cos -e ec i eness o heal hca e se ices, can
p o ide no el insigh s o designing heal h in e en ions and deepen ou unde s and-
ing o he in e connec ions be ween heal h and labo ma ke ou comes (Böcke man
and Ilmakunnas 2009; Picchio and Ubaldi 2024). This pe spec i e is pa icula ly el-
e an o cance s, which a e a majo con ibu o o he global disease bu den, wi h
hei impac expec ed o inc ease in he coming decades due o aging popula ions
(Global Bu den o Disease 2019 Cance Collabo a ion 2022). The e o e, he well-
being impac s o cance a e policy- ele an globally.
The e a e many possible easons o indi ec e ec s. Fi s , he heal h shock may
cause he pe son alling ill o educe hei e ec i e labo supply o e en wi hd aw
comple ely om he labo ma ke . I has p e iously been demons a ed ha a cance
diagnosis can lead o signi ican economic losses in households ac oss a ious ins i-
u ional con ex s (B adley e al. 2005; Jeon 2017; Vaala uo 2021). This can dis u b
he p e-exis ing a angemen on he join labo supply o spouses, especially when he
pe son alling ill is he main b eadwinne in he amily.1Consequen ly, he una ec ed
spouse may eac o his by inc easing he /his labo supply o main ain he amily’s
income le el and ma e ial well-being, in he spi i o wha has o en been called he
“added wo ke hypo hesis” (Mince 1962) o “added wo ke e ec ” (Lundbe g 1985).
Limi ed social secu i y, subs an ial deb , es ic ed access o c edi , and high heal h-
ca e cos s can signi ican ly exace ba e his e ec . The second po en ial d i e o he
indi ec labo ma ke e ec s leads o he opposi e consequences. The heal hy spouse
migh educe he /his labo supply due o he p o ision o ca e (known as he “ca egi e
e ec ”) o conce ns o he physical/men al well-being o a close amily membe and
a desi e o mo e sha ed leisu e ime ( he “ amily e ec ”).2
Beyond a ec ing labo supply, a cance diagnosis can cause psychological dis ess
among amily membe s. The unce ain y abou ecu ence, su i al, and e u ning o
wo k can signi ican ly impac no jus he pa ien bu also hei lo ed ones, leading o
las ing e ec s on hei li es (Mellon e al. 2007; Guan e al. 2021). This dis ess may
ex end o heal hy spouses as well, as some e idence sugges s (Hu e al. 2023; Angelini
and Cos a-Fon 2023). Mo eo e , cance , as a sudden, se e e, and unexpec ed shock
o he amily, can signi ican ly a ec he ma i al s abili y o couples na iga ing his
di icul si ua ion (Syse 2008).
While amily spillo e e ec s ha e ecen ly a ac ed inc eased in e es among
esea che s and policy ad oca es in he ield, empi ical e idence on he opic emains
inconclusi e. The di e gen esul s o exis ing esea ch a e likely explained, in pa , by
di e ences in ins i u ional con ex s, da a cha ac e is ics, empi ical me hods applied,
and he speci ic heal h shocks examined. To u he complica e he pic u e, exis ing
pa e ns in he di ision o household labo supply may subs an ially a ec he esponse
o he una ec ed spouse (see also Riekho and Vaala uo 2021; Vaala uo e al. 2023).
The couple’s join pa e n o labo supply p io o he heal h shock a ec s he need,
oppo uni ies, and obs acles o adjus ing labo ma ke pa icipa ion. In addi ion o labo
1Th oughou he pape “spouse” e e s o a signi ican o he in a ma iage o cohabi a ion.
2Fo a de ailed discussion o he ele an e minology, see Bobinac e al. (2010).
123
A amily a ai ? Long- e m economic and men al heal h… Page 3 o 30 19
ma ke esponses, hese ea u es may also in luence o he dimensions o well-being,
including psychological heal h and ma i al s abili y.
In his a icle, ou p incipal esea ch ques ion is how a spouse’s cance a ec s
he labo ma ke ou comes and men al heal h o he heal hy spouse and he couple’s
ma i al s abili y. Fu he mo e, while labo ma ke esponses ha e been epo ed in he
p io li e a u e, we con ibu e by in es iga ing he he e ogeneous esponses by ela i e
income posi ion wi hin he household. Iden i ying he causal e ec s o a spouse’s
cance diagnosis is challenging. Ou iden i ica ion s a egy elies on quasi- andom
a ia ion in he iming o he cance diagnoses wi hin he es ima ion window, using
a dynamic di e ence-in-di e ences o e en s udy design, as ou lined by Fadlon and
Nielsen (2019,2021). This app oach also enables us o assess he key iden i ica ion
assump ions o he empi ical speci ica ion.
Ou s udy ad ances he unde s anding o cance ’s b oade impac s on well-being
in h ee impo an ways. Fi s , we in eg a e heo ies and concep s o household di i-
sion o labo in o he empi ical models. We examine he he e ogeneous impac s by
he ela i e income s a us o he spouses p io o he cance diagnosis sepa a ely o
men and women. Rela i e income is ele an because i en ails in o ma ion on he
po en ial inancial losses o he su i ing spouse caused by cance , and signals po en-
ial economic independence wi hin he household. While some p e ious s udies ha e
examined gende di e ences in he spillo e e ec s, hey ha e no conside ed he e ec
o ela i e income s a us sepa a ely (independen ly o gende ) o ocused on long- e m
impac s. Second, ou s udy ocuses on he psychological spillo e e ec s o cance
and he impac s on ma i al s abili y. These wo dimensions o well-being ha e been
only a ely examined in he li e a u e, especially in connec ion o ela i e income s a-
us. Thi d, in con as o many o he s udies, we use pa icula ly long ollow-up ime
as some a ibu es associa ed wi h cance , e.g., unce ain y abou ecu ence, yea s o
su i al, and e u n o wo k, may ha e a p olonged e ec on he cance pa ien s’ li es
bu also on he li es o hei lo ed ones. Using he panel s uc u e o ou da a om
Finland o e he yea s 1995–2019, we ollow couples 5 yea s be o e and 10 yea s a e
he ini ial cance diagnosis.
While he e exis s only a nascen body o quasi-expe imen al esea ch on he indi-
ec e ec s o heal h shocks on he spouse’s labo ma ke ou comes, he connec ion
be ween heal h and labo supply wi hin he amily is no new o he ield o economics.
Al eady in he 1970s, Pa sons (1977) analyzed he impac o amily s uc u e on men’s
heal h, wo k hou s, and ea nings and obse ed ha poo heal h signi ican ly educed
men’s wo k hou s and ea nings bu did no no ably inc ease wo k hou s among o he
amily membe s. Be ge (1983) and Be ge and Fleishe (1984) also ound only small
inc eases in wi es’ wo k hou s and no subs an ial impac on labo o ce pa icipa ion
in esponse o husbands’ poo heal h.
In a mo e ecen and mo e closely ela ed s udy, Jeon and Pohl (2017) examined
he e ec o di e en cance diagnoses on he spouse’s employmen and ea nings
ajec o ies based on Canadian egis e da a. We complemen hei s udy by analyzing
he ole o ela i e income wi hin he household and by examining a wide se o
ou comes, including psycho opic d ug use. We a e no awa e o any p e ious quasi-
expe imen al s udies analyzing he spillo e e ec s o a heal h shock on men al heal h.
123
19 Page 4 o 30 P. Böcke man e al.
While some s udies (e.g., Bom e al. 2019; S öckel and Bom 2022; Angelini and
Cos a-Fon 2023) ha e examined men al heal h ou comes among spouses, hey ha e
no employed dynamic di e ence-in-di e ences o simila iden i ica ion s a egies o
ease ou causal e ec s.
Me hodologically, he closes o ou s udy is Fadlon and Nielsen (2021), who in es-
iga ed households’ labo supply esponses o a al and se e e non- a al heal h shocks
using Danish da a. While using a simila iden i ica ion s a egy, we concen a e on a
di e en heal h shock (i.e., cance ) and p o ide e idence also on he psychological
well-being e ec s. Mo eo e , we analyze he he e ogenei y o labo supply esponses
based on he ela i e income posi ion o couples and e alua e spillo e e ec s bo h in
he sho and long un.
Ou s udy is also ela ed o concu en wo k by A ie a and Li (2023) ha ocuses
on he e ec o eme gency depa men isi s (i.e., acu e heal h shocks equi ing u gen
ca e bu wi h a po en ially sho du a ion) on in a- amily adjus men o labo supply
and ca e in he U.S. con ex . O e all, we con ibu e o he eme ging li e a u e on he
opic by in es iga ing longe - e m impac s on labo supply and men al well-being as
well as he po en ial mechanisms behind he spousal e ec s in mo e de ail.
Ou indings om Finland, a comp ehensi e No dic wel a e s a e, a e likely o
illumina e he ins i u ional di e ences ha d i e labo supply esponses in a ious
coun y con ex s. Addi ionally, he Finnish con ex holds b oade in e es o wo o he
easons. Fi s , we examine he e ec s o cance on o al amily income, including
ecei ed social ans e s. This issue is highly policy- ele an in o he high-income
coun ies as hey de elop mo e comp ehensi e social sa e y ne s o amilies o ackle
he inancial bu den caused by ch onic illnesses. Second, Finland’s cance su i al
a es a e among he wo ld’s highes , which highligh s he impo ance o unde s anding
he indi ec e ec s o a heal h shock a he amily le el. In he nea u u e, he indi ec
labo ma ke e ec s o poo heal h migh become pa icula ly salien in aging socie ies
ha aim o p olong wo king ca ee s while econciling in o mal ca e and paid wo k
among olde employees.
We obse e ha emale spouses inc ease hei employmen o some yea s a e a
se e e heal h shock, which is consis en wi h he added wo ke e ec . Howe e , he
magni ude o he impac on annual ea nings is negligible. Among men, we obse e he
opposi e: male spouses’ ea nings dec ease once hei pa ne alls ill. O e all, cance
diagnosis causes a he small changes in he labo supply o spouses bu ela i ely
la ge inc eases in he use o psycho opic medica ion and psychia ic ou pa ien isi s.
Mo e impo an ly, ou esul s shed ligh on labo supply esponses wi hin amilies,
conside ing bo h he b eadwinne s a us and he su i al o he cance pa ien . We
obse e ha bo h men and women expe ience dec easing ea nings in cases o non-
a al cance s, wi h g ea e de ici s o seconda y ea ne s. Howe e , esponses in he
ex ensi e ma gin a e negligible. Bo h women and men mode a ely inc ease hei psy-
cho opic medica ion, wi h he mos signi ican ise obse ed among seconda y-ea ne
women. No ably, seconda y-ea ne women also expe ience a posi i e e ec on ma i-
al s abili y, while men and b eadwinne women emain una ec ed. In a al cance s,
a clea e ela ionship eme ges be ween ela i e income and labo supply decisions.
The ea nings esponses a e linea ly ela ed o he p e-cance income sha e wi hin
123
A amily a ai ? Long- e m economic and men al heal h… Page 5 o 30 19
he household, indica ing ha he g ea e he income sha e o he deceased spouse,
he g ea e he inc ease in he labo supply a he in ensi e ma gin o he su i ing
spouse. In he long- e m, seconda y-ea ne women also demons a e inc eased labo
supply a an ex ensi e ma gin. This implies ha he su i ing spouse compensa es
o he economic loss by adding labo supply. A he same ime, his g oup shows
he mos subs an ial inc ease in psycho opic medica ion use, sugges ing a connec ion
be ween he ma kedly inc eased labo supply and psychia ic symp oms om be ea e-
men . O e all, he ad e se men al heal h e ec s a e subs an ial bo h among men and
women, in bo h he sho - and long- e m. These indings highligh he impo ance o
he ollow-up du a ion, he su i al s a us o he ill spouse, and he b eadwinne s a us
wi hin he household in in luencing he esul s. Consequen ly, hey may help o esol e
some inconsis encies p esen in he exis ing e idence on he opic.
The a icle is s uc u ed as ollows. Sec ion 2o e s an o e iew o he ele an
li e a u e. Sec ion 3desc ibes he Finnish egis e da a and he empi ical amewo k.
Sec ion 4p esen s he es ima ion esul s. The inal sec ion o e s a comp ehensi e
discussion o he key indings.
2 Concep ual amewo k
Indi iduals conside he well-being and economic p ospec s o hei en i e household,
no jus hei own, when making labo supply decisions (Mince 1962; Blundell and
Walke 1982; Becke 1991). Ea ly in hei ela ionship, spouses o en nego ia e he
di ision o household labo , whe e adi ionally, he husband specializes in paid wo k
ou side he home, and he wi e, in unpaid household wo k a home (Becke 1991;
Lei a 1992).
The household ope a es as an economic uni , sha ing esou ces and isks. A heal h
shock, like cance , can signi ican ly change he household’s economic si ua ion,
impac ing labo supply, ma i al s abili y, and men al heal h, all c ucial o assess-
ing amily well-being. Exis ing esea ch on he spillo e e ec s o heal h shocks on
spouses has p ima ily s udied labo ma ke impac s, yielding inconclusi e esul s.
No ably, Coile (2004) examined hea a acks and new cance diagnoses among olde
adul s in he U.S., e ealing only a small added wo ke e ec o men and none o
women. On he o he hand, Jeon and Pohl (2017) obse ed a signi ican dec ease
in labo supply among Canadians whose spouses we e diagnosed wi h cance , pa -
icula ly among men a he in ensi e ma gin. The au ho s in e p e his inding as
indi iduals educing e ec i e labo supply o p o ide ca e o hei sick spouses and
o sha e leisu e ime. Simila ly, Anand e al. (2022) ound a educ ion in labo o ce
pa icipa ion among po en ial ca egi e s ollowing a spouse’s heal h shock. In con-
as , s udies like Giaquin o e al. (2022) using UK da a and Jolly and Theodo opoulos
(2023) wi h SHARE da a om Eu ope, show minimal changes in labo supply bu
highligh an inc eased ocus on ca egi ing and a highe likelihood o e i emen .
E idence on he long- e m spillo e e ec s o heal h shocks on psychological
well-being is spa se, and dynamic di e ence-in-di e ences designs ha e been a ely
u ilized in his esea ch. Recen s udies using su i al analysis and egis e da a om
123
19 Page 6 o 30 P. Böcke man e al.
Denma k and Sweden indica e ha a spouse’s cance diagnosis inc eases he isk o
ecei ing a psychia ic diagnosis in hospi al-based inpa ien o ou pa ien ca e com-
pa ed o ma ched con ols (Hu e al. 2023). Su ey e idence also suppo s he iew
ha cance ele a es he psychological dis ess o spouses. This inc eased dis ess may
s em om a g ea e ca egi ing bu den, impac ing he spouse’s men al heal h in a
dose- esponse manne (Bom e al. 2019; S öckel and Bom 2022), bu also because
he li es o spouses a e in ima ely linked in e ms o emo ional well-being and amily
esponsibili ies (No house and McCo kle 2015).
In a ecen s udy closely ela ed o ou s, Angelini and Cos a-Fon (2023)usea
c oss-sec ional su ey da a om Eu ope (SHARE) and ind ha a al cance is sig-
ni ican ly associa ed wi h he su i ing pa ne ’s well-being, leading o inc eased
dep ession, loneliness, and sleep p oblems. While being an impo an addi ion o
he heal h spillo e li e a u e, hei s udy depa s om he con en ional app oach by
using baseline cha ac e is ics and a i s di e ences speci ica ion ha may no ully
accoun o unobse ed ac o s in luencing changes o e ime. Ou a icle con ibu es
o p e ious esea ch by using mo e comp ehensi e egis e da a on heal h- ela ed con-
sump ion o psycho opic medicine and public psychia ic se ices wi hin a dynamic
di e ence-in-di e ences amewo k.
Gende di e ences in spousal esponses o heal h shocks, pa icula ly cance , a e
shaped by su i al a es and he ole o he p ima y b eadwinne , o en he male. The
ela i e income o each spouse has a signi ican impac on how households espond
o such shocks. I he p ima y b eadwinne alls ill, signi ican labo adjus men s
may be needed, po en ially leading o inc eased psychological s ess and changes in
amily dynamics (Becke e al. 1977). Howe e , when he seconda y-ea ne spouse is
a ec ed, labo adjus men s migh be smalle . Consis en wi h his iew, Fadlon and
Nielsen (2021) ound ha su i ing widows, bu no widowe s, inc eased hei labo
supply ollowing a al e en s, linking inancial loss o labo o ce pa icipa ion. This
inding indica es ha sel -insu ance migh play a c ucial ole in how amilies adjus
hei labo in esponse o heal h shocks.
Heal h shocks can a ec ma i al s abili y, especially agains he backd op o chang-
ing gende no ms in No dic coun ies, whe e women inc easingly academically
ou pe o m men, challenging he adi ional male b eadwinne model. S udies like
Be and e al. (2015) show ha emale b eadwinning can in luence ma i al sa is ac-
ion and s abili y, a end no ed in he U.S. in he 1960s and 1970s bu less so in he
1990s (Schwa z and Gonalons-Pons 2016), and is associa ed wi h inc eased ma i al
dissolu ion (Fos e and S a on 2021). This highligh s he need o explo e how gen-
de oles and ela i e household income in luence ma i al s abili y ollowing heal h
shocks.
Con a y o Becke e al. (1977), who sugges ed ha heal h issues could nega i ely
impac ma i al s abili y by a ec ing ai s like income po en ial and heal h, ecen s ud-
ies o e di e en iews. Fo example, Bünnings e al. (2021), analyzing Ge man da a,
disco e ed ha a spouse’s heal h decline does no necessa ily lead o ma i al ins abili y
and may e en s eng hen he ela ionship. Mo eo e , Ehle (2021) examined how a
heal h shock’s impac on ma i al s abili y a ies depending on he expec ed su i o ’s
pension, e ealing a posi i e co ela ion be ween po en ial su i o bene i s and he
123
A amily a ai ? Long- e m economic and men al heal h… Page 7 o 30 19
likelihood o s aying ma ied a e a heal h shock o emale pa ne s. This esul sug-
ges s ha highe economic dependency on he a ec ed spouse migh enhance ma i al
s abili y, as he una ec ed pa ne may ha e ewe al e na i es ou side he ma iage.
Con e sely, lowe economic dependence could ha e he opposi e e ec .
The exis ing li e a u e p o ides insigh s in o how he economic con ibu ions o a
pa ne , ela i e o hose o he a ec ed spouse, in luence adjus men s in labo supply
and ma i al s abili y. Howe e , he e ec s on men al heal h emain unexplo ed. Ou
s udy is he i s o examine he ole o ela i e income in mode a ing he men al heal h
e ec s o a spouse’s heal h shock. We examine wo hypo heses: Fi s , in households
whe e he una ec ed spouse con ibu es less economically, hey may ace g ea e
men al heal h challenges due o inancial s ess. Second, i he una ec ed spouse is
he p ima y ea ne , hey migh encoun e less inancial s ess bu mo e emo ional and
ca egi ing bu dens, impac ing hei men al heal h di e en ly.
3 Empi ical app oach
3.1 Resea ch design and iden i ica ion
Ou empi ical app oach employs he dynamic di e ence-in-di e ences design, simila
o Fadlon and Nielsen (2019,2021), he ea e e e ed o as FN DiD. By u ilizing his
iden i ica ion s a egy, we c ea ed coun e ac ual scena ios o couples in which one
spouse ecei ed a cance diagnosis. These we e de i ed om among couples who we e
diagnosed yea s la e . Employing households a ec ed by cance as he con ol g oup
is designed o educe he selec ion bias encoun e ed in s aigh o wa d case-con ol
compa isons.
Figu e 1illus a es ou esea ch design using examples whe e he ea men g oups
consis o indi iduals whose spouses we e diagnosed wi h cance in he yea 2000.
In Panel A, we plo he yea ly indica o o any psycho opic d ug pu chase o his
ea men g oup and compa e i o he ajec o y o he same ou come in indi iduals
whose spouses we e diagnosed wi h a cance diagnosis 11 yea s la e (=11).
The ollow-up con inues un il he las yea when he con ol g oup has no ye been
ea ed, i.e., yea 2010. Al hough he p e-e en le els o he wo g oups a e e y
simila , di ec compa ison may no be app op ia e due o di e ences in age and sex
dis ibu ions be ween he g oups. To ensu e a alid compa ison, we weigh ed he
ou comes in he con ol g oups acco ding o he age-sex dis ibu ion o he ea men
g oup, e ec i ely mimicking ma ching along hese dimensions. Panel B demons a es
ha weigh ing sligh ly al e s he ou come le el o he con ol g oup once age and sex
a e adjus ed. Bo h Panels A and B indica e ha he likelihood o psycho opic d ug
pu chases signi ican ly inc eases o indi iduals once hei spouse is diagnosed wi h
cance , compa ed o he con ol g oup.
In Panel C, we supplemen he compa ison wi h wo al e na i e con ol g oups: hose
a ec ed by a spouse’s cance diagnosis in 2006 (=6) and hose whose spouse ha e
no been diagnosed wi h cance . The ou come dynamics and le els in hese al e na i e
con ol g oups a e e y simila un il 2006. A e wa d, he g oup diagnosed wi h cance
123
19 Page 8 o 30 P. Böcke man e al.
A. Resea ch design wi h , unweigh ed
2000
2011
.06
.08
.1
.12
.14
.16
.18
.2
Psycho opic medica ion (pp.)
1995 2000 2005 2010
Yea
B. Resea ch design wi h , weigh ed
2000
2011
.06
.08
.1
.12
.14
.16
.18
.2
Psycho opic medica ion (pp.)
1995 2000 2005 2010
Yea
C. Compa ison wi h al e na i e con ol g oups
No diagnosis
2000
2006
2011
.06
.08
.1
.12
.14
.16
.18
.2
Psycho opic medica ion (pp.)
1995 2000 2005 2010
Yea
D. Cons uc ion o coun e ac ual
2011
2000
Coun e ac ual
.06
.08
.1
.12
.14
.16
.18
.2
Psycho opic medica ion (pp.)
-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
Yea s since cance diagnosis
Fig. 1 Illus a ions o he FN DiD design wi h psycho opic medica ion pu chases. Panel A compa es he
ea men g oup wi h coho s ha ecei ed a cance diagnosis 11 yea s la e . Panel B includes he esponses
o households ha unde wen he same shock in di e en yea s, alongside ma ched con ols ha ha e no
his o y o cance diagnosis. Panel C demons a es he cons uc ion o he coun e ac ual, based on he p e-
cance di e ence in ou comes. The ou comes o po en ial con ol g oups a e adjus ed based on he age and
sex dis ibu ion o he ea men g oup
in 2006 becomes “ ea ed” and is no longe iable as a con ol g oup o s udying he
longe - e m impac s o cance .
The ocus on he economic impac s o cance ypically spans 3–5 yea s in ela ed
s udies (e.g., Jeon 2017; Jeon and Pohl 2017,2019; Vaala uo 2021). Howe e , Tai e al.
(2005) a gue ha o ce ain cance s, like panc ea ic and s omach cance s, a pla eau in
s a is ical cu e a es is obse ed wi hin 10 yea s. Ye , o slowe p oli e a ing cance s,
such as hy oid and ea ly b eas cance s, eaching his pla eau migh ake e en decades.
We a gue ha a ime ame sho e han 10 yea s migh o e look signi ican long-
e m impac s o cance wi hin he amily. The cou se o he illness, la e ea men
e ec s, and possible ecu ence can ha e endu ing emo ional impac s on amily mem-
be s (Mellon e al. 2007; Guan e al. 2021). Addi ionally, he e ec s o heal h shocks
on amilies o en pe sis beyond 4 o 5 yea s a e he e en .3The e o e, a 10-yea
pos -cance ollow-up pe iod is jus i ied.
This decision ules ou he con ol g oup diagnosed wi h cance in 2006 as a po en ial
con ol g oup, lea ing =11 and una ec ed ma ched con ol g oup as po en ial
candida es o cons uc ing he coun e ac ual. We op ed o FN DiD ollowing ecen
3Fo example, Fadlon and Nielsen (2019) demons a e ha a heal h shock in one amily membe a ec s
he heal h beha io s o o he s o mo e han 4 yea s pos -shock, and Vaala uo e al. (2023) ind ha he
ad e se impac o child cance on a mo he ’s ea nings does no dissipa e wi hin 5 yea s pos -shock.
123
A amily a ai ? Long- e m economic and men al heal h… Page 15 o 30 19
Table 1 con inued
Men (wi e has cance ) Women (husband has cance )
Va iable Con ol g oup T ea men g oup Adj. Di e ence P Con ol g oup T ea men g oup Adj. Di e ence P
Psychia ic ou pa ien isi s 0.008 0.009 0.001 0.254 0.013 0.014 0.002 0.127
Psycho opic medica ion 0.085 0.093 0.000 0.871 0.136 0.156 0.007 0.018
An ipsycho ic medica ion 0.009 0.010 0.000 0.938 0.013 0.016 0.002 0.077
Anxioly ic medica ion 0.022 0.025 0.001 0.631 0.035 0.038 0.000 0.902
Hypno ic/seda i e medica ion 0.027 0.032 0.000 0.948 0.043 0.056 0.005 0.005
An idep essan medica ion 0.044 0.048 0.001 0.512 0.082 0.092 0.005 0.062
Opioid medica ion 0.020 0.021 −0.001 0.240 0.019 0.021 0.000 0.930
Cha lson como bidi y index 0.110 0.114 −0.008 0.027 0.105 0.125 0.008 0.038
Spouse dies wi hin 10 yea s 0.000 0.255 0.254 <0.001 0.000 0.437 0.438 <0.001
N 34484 23298 35849 20264
No es: This able p esen s sample means o he ea men and he con ol g oups, b oken down sepa a ely o men and women, o he yea p eceding he index diagnosis.
Columns 3 and 7 epo he age-adjus ed mean di e ence in he backg ound cha ac e is ics be ween he ea men and con ol g oups, and columns 4 and 8p esen he
co esponding p- alues. The sample comp ises adul s aged 28–64 whose spouses we e diagnosed wi h cance in Finland du ing he pe iods 2000 o 2008 ( ea men g oup)
and 2011 o 2019 (con ol g oup)
123
19 Page 16 o 30 P. Böcke man e al.
indi idual, i.e., he b eadwinne s a us. Following Bünnings e al. (2021); Fos e and
S a on (2021), we de ine a b eadwinne as a spouse who ou -ea ns he o he . We
cons uc ed a bina y a iable ha assumes a alue o 0 i he indi idual’s income
con ibu ion sha e was below 50% (indica ing he seconda y ea ne ) 1-yea p io o he
index diagnosis and 1 i he sha e was abo e 50% (indica ing he main b eadwinne ).
Table 1desc ibes he s udy sample, compa ing he ea men and con ol g oups
be o e he index diagnosis. The index diagnosis deno es he yea o cance diagnosis
o ea ed indi iduals and a placebo diagnosis yea o he con ol g oup, occu ing
11 yea s p io o he ac ual diagnosis yea o he con ol g oup. Men (husbands o
he cance pa ien ) we e p edominan ly in he b eadwinne ca ego y (72%), while
women (wi es o he cance pa ien ) we e less ep esen ed (31%). No able di e ences
be ween he g oups include employmen s a us, e i emen p obabili y, and heal h, as
measu ed by he Cha lson como bidi y index (Cha lson e al. 1987) using hospi al da a
om 1996 un il he yea p eceding he index diagnosis. These dispa i ies a e la gely
due o he di e en age dis ibu ions be ween he g oups. Adjus ing o a bi h yea
educes he a e age di e ences in he backg ound cha ac e is ics, hough some small
di e ences emain, such as in ea nings, disposable income, he likelihood o e ia y
educa ion, and heal h. These a ia ions jus i y he use o di e ence-in-di e ences
es ima ion wi h indi idual ixed e ec s.
Cance - ela ed mo ali y educed he household size by one in 34% o amilies
du ing he ollow-up pe iod wi hin he ea men g oup. No ably, men who we e diag-
nosed wi h cance had a conside ably highe likelihood o dea h (44%) compa ed o
women (26%) du ing he 10-yea pos -cance ollow-up pe iod.
4 Resul s
4.1 Main e ec s o spousal cance
We s a ou empi ical analysis by p esen ing he o e all esul s based on Eq. 1.
The es ima es om his speci ica ion, accompanied by he co esponding 95% con i-
dence in e als a e depic ed g aphically in Fig. 2, while he pa ame e coe icien s a e
epo ed in Appendix Tables A1–A4.
The igu es plo he change in he ou comes o in e es ela i e o he yea be o e
spousal cance diagnosis in he ea men g oup. Impo an ly, he igu es do no exhibi
clea p e- ends, he eby suppo ing he key iden i ica ion assump ion unde lying ou
empi ical speci ica ion.
Fo women, we ind e idence ha hey educe hei employmen in he sho -
e m bu inc ease employmen in he long- e m a e hei spouse’s cance diagnosis
(Panel A o Fig. 2). The a e age di e ence-in-di e ences es ima e o women is ze o
(Appendix Table A2). Among male spouses, he e ec was indis inguishable om
ze o h oughou he ollow-up. O e all, he labo ma ke impac s a e close o ze o,
which in con as o he esul s in Jeon and Pohl (2017) o Canada acco ding o which
spouses o he cance pa ien s dec ease subs an ially hei labo supply (abou 2 o 3
pp.) a he ex ensi e ma gin leading o a lowe le el o ea nings. Mo eo e , he changes
123
A amily a ai ? Long- e m economic and men al heal h… Page 17 o 30 19
A. Employmen
-2
-1
0
1
2
E ec on employmen (pp.)
-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
Yea s since diagnosis
Men (wi e has cance ) Women (husband has cance )
B. Re i emen
-3
-2
-1
0
1
E ec on e i emen (pp.)
-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
Yea s since diagnosis
Men (wi e has cance ) Women (husband has cance )
C. Ea nings
-3
-2
-1
0
1
2
E ec on ea nings (%)
-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
Yea s since diagnosis
Men (wi e has cance ) Women (husband has cance )
D. HH disposable income
-15
-10
-5
0
E ec on HH disp. income (%)
-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
Yea s since diagnosis
Men (wi e has cance )
Women (husband has cance )
E. Psycho opic d ug use
-2
0
2
4
E ec on psycho opic medica ion (pp.)
-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
Yea s since diagnosis
Men (wi e has cance ) Women (husband has cance )
F. Psychia ic ou pa ien isi s
-.5
0
.5
1
E ec on psychia ic ou pa ien isi (pp.)
-5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10
Yea s since diagnosis
Men (wi e has cance ) Women (husband has cance )
Fig. 2 The e ec o spousal cance on labo supply, income, psycho opic d ug use, and psychia ic ou pa-
ien isi s. The poin es ima es, wi h he shaded a eas indica ing he 95% con idence in e als, ep esen
he di e ences in ou comes compa ed o he con ol g oup ela i e o he yea p eceding he index cance
diagnosis. The con ol g oup is composed o indi iduals diagnosed wi h cance 11 yea s la e ela i e o he
ea men g oup. The e ical line a ime =−0.5 depic s he momen o he cance diagnosis. S anda d
e o s a e clus e ed a he indi idual le el. The co esponding e en s udy es ima es a e de ailed in Tables
A1–A4
123
19 Page 18 o 30 P. Böcke man e al.
in employmen o women we e e lec ed in he p obabili y o e i ing om he labo
ma ke (Panel B). Consis en wi h he employmen e ec s, he e ec on he p obabili y
o e i ing was la ge o women compa ed o men in he long- e m. While his esul
con as s wi h ha o ega ding No dic coun ies in Jolly and Theodo opoulos (2023),
i is in acco dance wi h a nega i e e ec on (ea ly) pension bene i s among widows
ound by Fadlon and Nielsen (2021). As he social insu ance s uc u e is e y simila
be ween Denma k and Finland, and da a and empi ical s a egy a e almos iden ical,
he conco dan esul s wi h Fadlon and Nielsen (2021) a e unsu p ising.
Howe e , when i comes o annual ea nings (Panel C), he magni ude o he e ec
was negligible o women. Addi ionally, o male spouses, hei ea nings dec eased
a e hei spouse ell ill. This inding is consis en wi h he amily e ec , al hough he
e ec was modes wi h less han a 2% educ ion in ea nings.
The es ima es o he household’s disposable income (Panel D) (as well as o
household’s o al income and household disposable income adjus ed by using OECD
modi ied equi alence scale shown in Appendix Tables A1 and A2) e eal a meaning ul
dec ease, om 10 o 15% in he medium o long un among emale spouses and
app oxima ely 5% among he male spouses. This esul e eals ha a emale spouse’s
inc ease in ea nings does no compensa e o he loss o he sick spouse’s income
wi hin he household. Ou esul s highligh ha inancial consequences ollowing a
spouse’s dea h end o be ha she o women, which migh explain he inding ha
cance leads o a la ge dec ease in owne -occupancy in housing among women bu
no among men (Appendix Tables A1 and A2). Women may need o liquida e hei
asse s o suppo household income and main ain ma e ial well-being.
Impo an ly, in addi ion o consequences in he labo ma ke , cance also a ec s he
spouse’s men al well-being. Acco ding o Panel E o Fig. 2, he p obabili y o using
psycho opic medica ion ini ially inc eased by app oxima ely 4 pe cen age poin s (pp.)
o women and abou 2 pp. o men, and he impac s s abilized a a ound 1.5–2 pp. o
bo h sexes. Rela i e o he baseline p obabili y o psycho opic medica ion use, he
ela i e inc ease in psycho opic medica ion was 13.0% o men and 14.1% o women
on a e age du ing he ull ollow-up pe iod. This esul is in line wi h he e idence
om Sweden and Denma k, sugges ing a 13% isk inc ease in psychia ic diso de s
ollowing a spouse’s cance (Hu e al. 2023). Appendix Tables A3 and A4 also epo
he co esponding es ima es ega ding sub-ca ego ies o psycho opic medica ion such
as an ipsycho ic, anxioly ic, an idep essan , and sleep (hypno ics/seda i es) medicine.
O e all, he dynamics o he e ec s a e la gely simila o psycho opics in gene al.
These ables also epo inc eases in he p obabili y o isi ing psychia is s in special-
ized public heal h ca e. This ou come indica es mo e se e e psychia ic symp oms.
The inc ease is la ge o women (DD es ima e 0.3 pp. s. baseline 1.1 pp.) han o
men (DD-es 0.2 pp. s. baseline 0.8 pp.). He ea e we epo psychia ic impac s
only in (any) psycho opic medica ion.
A po en ial conce n ega ding he alidi y o he es ima es is he endogenous na u e
o cance . The e en s udy speci ica ion ensu es ha he compa isons a e conduc ed
o indi iduals o he same sex, age, and cance ype o he spouse, and educa ion le el
bu wi h he iming di e ence o he spouse’s cance diagnosis o 11 yea s. The main
conce n he e o e is ela ed o he iming o he diagnosis. The iming di e ence can
123
A amily a ai ? Long- e m economic and men al heal h… Page 19 o 30 19
po en ially e eal di e ences be ween household li ing condi ions and hei heal h
beha io s. Hence, as a obus ness check, we e-es ima ed he e ec s o spousal cance
using only a subse o cance diagnoses ha a e less ela ed o heal h beha io s. This
subse o cance s includes (ICD-10 ca ego yin pa en heses): Gallbladde cance (C23),
B eas cance (C50), and O a ian cance (C56), P os a e cance (C61), Tes icula
cance (C62), Thy oid cance (C73), Myeloma (C90), Non-Hodgkin lymphoma (C82-
85,C96), Leukaemia (C91-C95), and B ain and o he cen al ne ous sys em cance s
(C70-72). The selec ion o cance s was based on he B i ish (B own e al. 2018) and
Aus alian (Wilson e al. 2018) es ima es o he ac ion o cance s ha a e p e en able
wi hin each cance diagnosis ca ego y. In ou analysis, we adop ed a conse a i e
app oach and included only hose ypes o cance es ima ed o be p e en able by up
o 30%. Encou agingly, he esul s based on his subsample (Appendix Tables A5 and
A6) a e by and la ge e y simila o he baseline esul s p esen ed in Fig. 2.Asan
addi ional obus ness check, we also es ic ed he sample o couples who had li ed
oge he o he en i e 5-yea pe iod and ound ha esul s we e quan i a i ely e y
simila (Appendix Tables A7 and A8).
Mo eo e , we examined whe he he use o an e en s udy app oach a ec s he
in e p e a ion o he main esul s. Appendix Fig. A4 shows he esul s o his exe cise
o ou p e e ed choice, FN DiD wi h a con ol g oup a ec ed by cance 6 yea s la e ,
using a s acked e en s udy design and ma ching wi h una ec ed households. We ind
ha he es ima es a e quali a i ely he same ac oss all ou app oaches and, o he
mos pa , also quan i a i ely e y simila .
As p e iously s a ed, he cance su i al a es di e ed no ably by gende . 44% o
men and 26% o women died du ing he 10-yea ollow-up (Table 1). This could lead
o po en ial di e ences in he spousal labo supply esponses be ween gende s. Fo
his eason, we nex p oceed o es ima e he e ec s sepa a ely by he su i al o he
cance pa ien .
4.2 He e ogenei y by b eadwinne s a us in non- a al cance s
We examine whe he he p e-cance ela i e income s a us wi hin he household in lu-
ences he impac s o spousal cance in non- a al cance s. We sepa a ely es ima ed
he b eadwinne he e ogenei y e ec s o emale and male spouses using Eq. 2on
ea nings, employmen , and psycho opic d ug use in he sho -, medium- and long-
e m. These esul s a e p esen ed in Table 2. The impac es ima es o employmen ,
psycho opic d ug use, and ma i al s a us a e p esen ed in pe cen age poin s and as a
pe cen age ela i e o he baseline alues. The p e-cance mean o he ou come wi hin
he b eadwinne s a us is epo ed in he igh mos column.
We ind ha he e is a educ ion in ea nings o seconda y ea ne s bo h o men
and women in non- a al cance s. In seconda y-ea ne women, ea nings dec ease by 3
% h oughou he ollow-up pe iod. In con as , seconda y-ea ne men expe ience an
ini ial dec ease o 2%, ollowed by income de ici s o 4% and 3% in he medium- and
long- e m, espec i ely. The poin es ima es a e nega i e also o b eadwinne s bu hey
a e no s a is ically signi ican . The di e ence in ea nings esponses by b eadwinne
s a us is s a is ically signi ican o women in he sho - e m and sugges i ely di e en
123
19 Page 20 o 30 P. Böcke man e al.
Table 2 E ec s o spousal cance by b eadwinne s a us in non- a al cance s
B eadwinne s a us Sho - e m Medium- e m Long- e m Con ol g oup mean
Es . % Es . % Es . %
A. Women (husband has cance )
Ea nings (%) B eadwinne −0.6[0.8]−0.7[1.1]−2.7[1.4]27549.1
Seconda y ea ne −2.8[0.7]*−3[0.9]−3.3[1.1]16880.1
Employmen (pp.) B eadwinne 0 [0.5]0−0.2[0.6]−0.3−0.4[0.7]−0.6 72.3
Seconda y ea ne −1.1[0.4]†−1.8−0.3[0.4]−0.6−0.2[0.5]−0.459.2
Psycho opic B eadwinne 0.7 [0.5]4.7 −0.1[0.5]−0.40.4[0.6]2.7 15.4
d ug use (pp.) Seconda y ea ne 1.5 [0.3]8.7 1.3 [0.3]*7.21.1[0.4]6.4 17.4
Ma ied wi h B eadwinne 0.2 [0.3]0.3 0 [0.4]0−0.2[0.5]−0.3 84.7
cance pa ien (pp.) Seconda y ea ne 0.6 [0.2]0.6 0.9 [0.3]†1 0.9[0.3]†1.1 89
B. Men (wi e has cance )
Ea nings (%) B eadwinne 0.1 [0.4]−0.7[0.6]−1.5[0.8]36195.2
Seconda y ea ne −1.8[1.3]−3.6[1.5]†−3.3[1.7]15184
Employmen (pp.) B eadwinne −0.2[0.3]−0.3−0.3[0.3]−0.5−0.1[0.4]−0.171
Seconda y ea ne −0.1[0.5]−0.10[0.6]0−0.2[0.7]−0.4 52.2
Psycho opic B eadwinne 0.6 [0.2]5.6 0.4 [0.2]3.5 0.6 [0.3]6.2 10.2
d ug use (pp.) Seconda y ea ne 0.6 [0.4]4.4 0.7 [0.4]5.3 0.8 [0.5]5.9 13.2
Ma ied wi h B eadwinne 0 [0.2]0−0.2[0.3]−0.20[0.3]087.1
cance pa ien (pp.) Seconda y ea ne 0.3[0.3]0.4 0.6[0.4]0.7 0.1[0.5]0.1 83.2
No es: Sho -, medium-, and long- e m impac s o spousal cance by di e en b eadwinne s a us in non- a al cance s. S anda d e o s (clus e ed a indi idual le el) a e
epo ed beside he poin es ima es in pa en heses. Sho - e m e e s o DD es ima es using pos -e en pe iods 0–2, medium- e m o pe iods 3–5, and long- e m o pe iods
6–10. Symbols †, * and ** e e o s a is ical signi icance 10%, 5% and 1% o he poin es ima es ela i e o he e e ence g oup (B eadwinne ). All es ima es a e based on
he iple-di e ence models p esen ed in Eq. 2
123
A amily a ai ? Long- e m economic and men al heal h… Page 21 o 30 19
o men in he medium- e m. In e es ingly, he esponses a e conside ably mo e sub le
in e ms o employmen , being nega i e only o seconda y-ea ne women in he sho -
e m. This sugges s ha in non- a al cance s, he spouses o he pa ien s adjus hei
labo supply mo e on he in ensi e ma gin han on he ex ensi e ma gin, in line wi h
he e idence p esen ed by Fadlon and Nielsen (2021).
The pu chases o p esc ibed psycho opic medica ions inc ease o bo h men and
women. The e ec s a e mos p onounced in seconda y ea ne s, wi h a 1.5 pe cen age
poin (pp) inc ease in he sho - e m, and 1.3 pp. and 1.1 pp. inc eases in he medium
and long- e m, espec i ely. He e ogenei y in ela i e income is s a is ically signi ican
only o women in he medium- e m.
A po en ial ac o a ec ing economic and men al heal h esponses is union s abili y.
In he inal panels o Table 2, we epo he impac o cance on being ma ied wi h
he cance pa ien . App oxima ely 87% o he couples we e ma ied a he baseline.
We ind ha cance has a posi i e e ec on ma iage o he seconda y ea ne (cance
pa ien is he b eadwinne ) and no e ec on he b eadwinne (cance pa ien is he
seconda y ea ne ). This esul is in line wi h p io esea ch by Ehle (2021) ha ound
ha heal h shock inc eases he p obabili y o ma iage among cohabi ing couples mo e
he highe a e expec ed su i o pension o he widow. Howe e , no such e ec s a e
ound o men.
4.3 He e ogenei y by b eadwinne s a us in a al cance s
Finally, we in es iga e whe he he impac o a cance diagnosis esul ing in dea h
wi hin he 10-yea ollow-up pe iod di e s by b eadwinne s a us. In his analysis, we
exclude ma i al ou comes, as dea h e mina es ma iage in he ea men g oup unless
i had al eady dissol ed be o ehand. The analysis is based on a sample om which we
exclude households whe e he cance pa ien did no die du ing he ollow-up pe iod.
These esul s a e epo ed in Table 3.
We ind ha a al spousal cance ini ially dec eases ea nings by 3% and 2% o
seconda y and b eadwinne women, espec i ely. In he long- e m, he e is a consid-
e able con as in he esponse in e ms o ela i e income. Ea nings o b eadwinne
women emained unchanged, whe eas hose o he seconda y ea ne s exhibi ed a 6%
inc ease. A simila pa e n is obse ed a he ex ensi e ma gin; b eadwinne s showed
no change in employmen , while seconda y ea ne s expe ienced a signi ican 4 pp.
inc ease in he long- e m. The labo ma ke esponses we e s a is ically signi ican ly
di e en o he seconda y ea ne and b eadwinne women only in he long- e m.
Howe e , he inc ease in he p obabili y o psycho opic medica ion pu chases was
consis en ly highe (1.4 pp. o 1.9 pp. highe ) o he seconda y ea ne s compa ed o
b eadwinne women.
The b eadwinne he e ogenei y was ma kedly di e en o men. Fo men, we ound
no s a is ically signi ican di e ences by he b eadwinne s a us o he mos pa ,
bu he coe icien s poin owa ds s a is ically signi ican esponses in he long- e m
wi h b eadwinne men dec easing and seconda y ea ne s inc easing hei labo supply
a he in ensi e ma gin. Addi ionally, b eadwinne men showed highe inc eases in
123
19 Page 22 o 30 P. Böcke man e al.
Table 3 E ec s o spousal cance by b eadwinne s a us in a al cance s
B eadwinne s a us Sho - e m Medium- e m Long- e m Con ol g oup mean
Es . % Es . % Es . %
A. Women (husband has cance )
Ea nings (%) B eadwinne −2[0.8]−0.1[1.1]−0.2[1.3]27398.9
Seconda y ea ne −2.7[0.8]1.8[1]6.2[1.2]** 16709.7
Employmen (pp.) B eadwinne −0.3[0.6]−0.50.5[0.7]0.7 0 [0.7]072
Seconda y ea ne −0.3[0.4]−0.61.7[0.5]2.9 4.3[0.6]** 7.3 58.8
Psycho opic B eadwinne 6.4[0.6]41.7 4.2[0.6]27.5 2.7[0.6]17.4 15.3
d ug use (pp.) Seconda y ea ne 7.8[0.4]* 44.8 6.1[0.4]** 35.4 4.3[0.5]* 24.5 17.3
B. Men (wi e has cance )
Ea nings (%) B eadwinne −1.9[0.7]−3.1[0.9]−2.3[1.1]36131.4
Seconda y ea ne −2.5[2]−2.1[2.2]2.2[2.5]† 15116.9
Employmen (pp.) B eadwinne −0.7[0.5]−1−0.6[0.5]−0.8 0 [0.6] 0 70.9
Seconda y ea ne 0.7[0.8]1.3 0.4[0.9]0.7 1.1[1.1]2.1 52.1
Psycho opic B eadwinne 4.3[0.4]42.2 4 [0.4]39.3 3.8[0.5]37.2 10.2
d ug use (pp.) Seconda y ea ne 2.9[0.6]* 21.7 4.6[0.7]34.7 3.3[0.8]24.9 13.2
No es: This able p esen s he sho -, medium-, and long- e m impac s o spousal cance by di e en b eadwinne s a uses in a al cance s. S anda d e o s, clus e ed a
indi idual le el, a e p esen ed alongside he poin es ima es in pa en heses. Sho - e m e e s o DD es ima es o pos -e en pe iods 0–2, he medium- e m o pe iods 3–5,
and he long- e m o pe iods 6–10. Symbols †, * and ** deno e s a is ical signi icance 10%, 5%, and 1% le els, espec i ely, o he poin es ima es ela i e o he e e ence
g oup (B eadwinne ). All es ima es a e de i ed om he iple-di e ence models p esen ed in Eq. 2
123
A amily a ai ? Long- e m economic and men al heal h… Page 23 o 30 19
psycho opic medica ion pu chases in he sho - e m compa ed o seconda y ea ne s,
bu no in he medium- and long- e m.
O e all, we ind ha dea h is he p ima y d i e behind he inc eased psycho opic
d ug use among spouses o cance pa ien s. When he cance pa ien su i es, spouses
a e ela i ely una ec ed, bu a al cance leads o no able inc eases in psycho opic
d ug pu chases. While he es ima es a e no di ec ly compa able, ou esul s align
wi h hose o Angelini and Cos a-Fon (2023), which sugges ha a al cance leads o
a subs an ial inc ease in sel - epo ed dep ession symp oms, while non- a al cance s
exhibi mo e sub le changes in psychological symp oms among spouses.
4.4 Reconciling he e idence ac oss he ela i e income dis ibu ion
O e all, he esul s om he p e ious sec ion sugges ha he ole o ela i e ea nings
is s onge o women han o men in esponses o spousal cance . Howe e , a bina y
indica o o b eadwinne s a us, de e mined s ic ly by a 0.5 cu -o does no cap u e
he sub le esponses along he ela i e income spec um. To be e unde s and he ole
o ela i e ea nings in he esponses o spousal cance , we di ided he sample in o
i e equal-sized g oups by he p e-shock ea nings income sha e o he indi idual.
Fo illus a i e pu poses, we ocus solely on long- e m esponses, i.e., he impac s
o spousal cance on ou comes measu ed 6 o 10 yea s a e he cance diagnosis.
Essen ially, we conduc ed di e ence-in-di e ences analyses o es ima e he long- e m
impac s sepa a ely o each ela i e income quin ile and by sex.
Figu e 3shows he poin es ima es along wi h hei 95% con idence in e als. On he
x-axis, he income quin ile sha es ep esen he spouse’s sha e o he o al household
income 1 yea p io o he diagnosis. The mean alues o income sha es wi hin each
quin ile a e epo ed in he pa en heses.
We ind ha he lowe a spouse’s ela i e income sha e wi hin he household,
he la ge he inc ease in hei ea nings. The ela ionship is nea ly linea in e ms
o ea nings o bo h men and women (as shown in Panels A and B). This sugges s
ha g ea e income losses due o losing a spouse co ela e wi h a la ge inc ease in
labo supply a he in ensi e ma gin, ela i e o he coun e ac ual ajec o y. This
esul aligns wi h he e idence om Fadlon and Nielsen (2021), which shows ha
he amoun o a spouse’s income los due o a al ca dio ascula e en s is posi i ely
linked o labo supply esponses among su i ing spouses. No ably, ou analysis also
sugges s ha his pa e n exis s in bo h a al and non- a al cance s.
The he e ogenei y in he esponses a he ex ensi e ma gin o labo supply is less
e iden (Panels C and D). We ind ha a spouse’s baseline income sha e is nega i ely
co ela ed wi h he added wo ke e ec bu only o women in a al cance s. Fo
bo h men and o women wi h non- a al cance s, ela i e income does no appea
o in luence employmen esponses. Collec i ely, hese esul s sugges ha men a e
mo e likely o adjus hei labo supply a he in ensi e ma gin, whe eas o women,
he ex ensi e ma gin accoun s o a la ge sha e o he changes in ea nings ollowing
a spouse’s cance diagnosis.
123
19 Page 24 o 30 P. Böcke man e al.
A. Women’s ea nings
-12
-6
0
6
12
18
24
E ec on ea nings (%)
1
(0.225)
2
(0.400)
3
(0.493)
4
(0.587)
5
(0.764)
Baseline (own) income sha e quin ile
Non- a al cance
Fa al cance
B. Men’s ea nings
-12
-6
0
6
12
18
24
E ec on ea nings (%)
1
(0.225)
2
(0.400)
3
(0.493)
4
(0.587)
5
(0.764)
Baseline (own) income sha e quin ile
Non- a al cance
Fa al cance
C. Women’s employmen
-6
-4
-2
0
2
4
6
E ec on employmen (pp.)
1
(0.225)
2
(0.400)
3
(0.493)
4
(0.587)
5
(0.764)
Baseline (own) income sha e quin ile
Non- a al cance
Fa al cance
D. Men’s employmen
-6
-4
-2
0
2
4
6
E ec on employmen (pp.)
1
(0.225)
2
(0.400)
3
(0.493)
4
(0.587)
5
(0.764)
Baseline (own) income sha e quin ile
Non- a al cance
Fa al cance
E. Women’s psycho opic
medica ion
-4
-2
0
2
4
6
8
E ec on psycho opic medica ion (pp.)
1
(0.225)
2
(0.400)
3
(0.493)
4
(0.587)
5
(0.764)
Baseline (own) income sha e quin ile
Non- a al cance
Fa al cance
F. Men’s psycho opic medica ion
-4
-2
0
2
4
6
8
E ec on psycho opic medica ion (pp.)
1
(0.225)
2
(0.400)
3
(0.493)
4
(0.587)
5
(0.764)
Baseline (own) income sha e quin ile
Non- a al cance
Fa al cance
G. Women’s ma iage
-4
-2
0
2
4
E ec on s aying ma ied (pp.)
1
(0.225)
2
(0.400)
3
(0.493)
4
(0.587)
5
(0.764)
Baseline (own) income sha e quin ile
Non- a al cance
H. Men’s ma iage
-4
-2
0
2
4
E ec on s aying ma ied (pp.)
1
(0.225)
2
(0.400)
3
(0.493)
4
(0.587)
5
(0.764)
Baseline (own) income sha e quin ile
Non- a al cance
Fig. 3 Spouses’ long- e m esponse es ima es along wi h 95% con idence in e als (yea s 6–10 a e s. 5
yea s p eceding he cance diagnosis) o a al cance , ca ego ized by hei p e-shock income sha e quin ile
wi hin he household’s o al ea nings. We di ided he sample in o i e equal-sized g oups based on he
spouses’ p e-cance sha e o he o al household income. Subsequen ly, we plo ed he a e age ou come
esponse agains he p e-cance mean income sha e o each quin ile
123
